EP3224603B1 - Device and method for material characterisation - Google Patents

Device and method for material characterisation Download PDF

Info

Publication number
EP3224603B1
EP3224603B1 EP15862380.1A EP15862380A EP3224603B1 EP 3224603 B1 EP3224603 B1 EP 3224603B1 EP 15862380 A EP15862380 A EP 15862380A EP 3224603 B1 EP3224603 B1 EP 3224603B1
Authority
EP
European Patent Office
Prior art keywords
detector
calibration
energy
pulse
spectrum
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
EP15862380.1A
Other languages
German (de)
English (en)
French (fr)
Other versions
EP3224603A4 (en
EP3224603A1 (en
Inventor
Paul Scoullar
Christopher Mclean
Shane Tonissen
Syed Khusro SALEEM
Brendan Allman
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southern Innovation International Pty Ltd
Original Assignee
Southern Innovation International Pty Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southern Innovation International Pty Ltd filed Critical Southern Innovation International Pty Ltd
Publication of EP3224603A1 publication Critical patent/EP3224603A1/en
Publication of EP3224603A4 publication Critical patent/EP3224603A4/en
Application granted granted Critical
Publication of EP3224603B1 publication Critical patent/EP3224603B1/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G01V5/228—
    • G01V5/224—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/00002—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for
    • H04N1/00007—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for relating to particular apparatus or devices
    • H04N1/00023—Colour systems

Definitions

  • This invention relates to a device for material identification of the invention, with particular application to inspection of freight or baggage for one or more types of material.
  • X-ray systems used in freight and baggage screening use a broad spectrum X-ray generator to illuminate the item to be screened.
  • a detector array on the opposite side of the item is used to measure the intensity of X-ray flux passing through the item.
  • Larger systems may have the option to have two or more X-ray sources so as to collect two or more projections through the cargo at the same time.
  • X-Ray screening systems use the differential absorption of the low energy and high energy X-rays to generate a very coarse classification of the screened material, and then use this coarse classification to generate a "false color" image for display. A small number of colors - as few as 3 in most existing systems - are used to represent material classification.
  • US 2013/315464 discloses a device for screening one or more items of freight or baggage comprising: a source of incident radiation configured to irradiate the one or more items; a plurality of detectors adapted to detect packets of radiation emanating from within or passing through the one or more items as a result of the irradiation by the incident radiation, each detector being configured to produce an electrical pulse caused by the detected packets having a characteristic size or shape dependent on an energy of the packets; one or more digital processors configured to process each electrical pulse to determine the characteristic size or shape and to thereby generate a detector energy spectrum for each detector of the energies of the packets detected, and to characterise a material associated with the one or more items based on the detector energy spectra.
  • a device for screening one or more items of freight or baggage comprising:
  • each packet of radiation is a photon and the plurality of detectors comprise one or more detectors each composed of a scintillation material adapted to produce electromagnetic radiation by scintillation from the photons and an pulse producing element adapted to produce the electrical pulse from the electromagnetic radiation.
  • the pulse producing element may comprise a photon-sensitive material and the plurality of detectors may be arranged side-by-side in one or more detector arrays of individual scintillator elements of the scintillation material each covered with reflective material around sides thereof and disposed above and optically coupled to a photon-sensitive material.
  • the scintillation material may be lutetium-yttrium oxyorthosilicate (LYSO).
  • the photon-sensitive material may be a silicon photomultiplier (SiPM).
  • SiPM silicon photomultiplier
  • the individual scintillator elements of one or more of the detector arrays may present a cross-sectional area to the incident radiation of greater than 1.0 square millimetre.
  • the cross-sectional area may be greater than 2 square millimetres and less than 5 square millimetres.
  • the one or more digital processors are further configured with a pileup recovery algorithm adapted to determine the energy associated with two or more overlapping pulses.
  • the one or more digital processors is configured to compute an effective atomic number Z for each of at least some of the detectors based at least in part on the corresponding detector energy spectrum.
  • the one or more digital processors may be configured to compute the effective atomic number Z for each of at least some of the detectors by: determining a predicted energy spectrum for a material with effective atomic number Z having regard to an estimated material thickness deduced from the detector energy spectrum and reference mass attenuation data for effective atomic number Z; and comparing the predicted energy spectrum with the detector energy spectrum.
  • the one or more digital processors may be configured to compute the effective atomic number Z for each of at least some of the detectors by: determining a predicted energy spectrum for a material with effective atomic number Z having regard to a calibration table formed by measuring one or more materials of known composition; and comparing the predicted energy spectrum with the detector energy spectrum.
  • the one or more digital processors is configured to perform the step of comparing by computing a cost function dependent on a difference between the detector energy spectrum and the predicted energy spectrum for a material with effective atomic number Z.
  • a gain calibration is performed on each detector individually to provide consistency of energy determination among the detectors and the one or more digital processors is further configured to calculate the detector energy spectrum for each detector taking into account the gain calibration.
  • a count rate dependent calibration is performed comprising adaptation of the detector energy spectra for a count rate dependent shift.
  • a system parameter dependent calibration is performed on the detector energy spectra comprising adaptation for time, temperature or other system parameters.
  • the one or more digital processors is further configured to reduce a communication bandwidth or memory use associated with processing or storage of the detector energy spectra, by performing a fast Fourier transform of the energy spectra and removing bins of the fast Fourier transform having little or no signal to produce reduced transformed detector energy spectra.
  • the one or more digital processors may be further configured to apply an inverse fast Fourier transform on the reduced transformed detector energy spectra to provide reconstructed detector energy spectra.
  • the one or more digital processors may be further configured with a specific fast Fourier transform window optimised to minimise ringing effects of the fast Fourier transform.
  • the one or more digital processors is further configured with a baseline offset removal algorithm to remove a baseline of a digital signal of electrical pulse prior to further processing.
  • the one or more digital processors is further configured to produce an image of the one or more items composed of pixels representing the characterisation of different parts of the one or more items deduced from the detector energy spectra.
  • the one or more digital processors is further configured to perform one or more of tiling, clustering, edge detection or moving average based on the effective atomic numbers determined for said plurality of detectors.
  • the one or more digital processors is further configured to perform threat detection based on one or more types of target material.
  • a method of screening one or more items of freight or baggage comprising the steps of:
  • the words "freight or baggage” encompass parcels, letters, postage, personal effects, cargo, boxes containing consumer or other goods and all other goods transported which are desirable or necessary to scan for certain types of materials, including but not limited to contraband, and dangerous or explosive materials which may be placed by accident or placed deliberately due to criminal, terrorist or military activity.
  • packets in relation to incident radiation includes individual massless quantum particles such as X-ray, gamma-ray or other photons; neutrons or other massive particles; and also extends in its broadest aspects to any other corpuscular radiation for which an energy of each corpuscle may be defined and detected.
  • energy spectrum in relation to a particular detector refers to a generation of energy values of the individual packets of radiation emanating from or passing through the part of the items under investigation as detected over a time interval from the particular detector, which energy values can comprise values over a range, typically continuous, and may be represented as a histogram of detection counts versus a plurality of defined energy bins, the number of bins representing the desired or achievable energy resolution and constituting at least 10 bins but preferably more than 50, 100 or 200 energy bins.
  • This invention relates to a method and apparatus for material identification using a range of radiation types for analysis.
  • the apparatuses and methods exemplified herein may be applied to X-ray screening, however, it will be appreciated that the apparatuses and methods could readily be modified for other types of incident radiation such as neutrons or gamma rays, or other types of emanating radiation, particularly by substituting a different form of detector unit, to detect for example electromagnetic, neutron, gamma-ray, light, acoustic, or otherwise. Such modifications are within the broadest aspect of the invention.
  • X-rays passing through matter interact with that matter via a number of modalities including: scattering off crystal planes, causing fluorescence X-ray emission from within the electron structure of the elements; and, scattering off nano-scale structures within the material being scanned.
  • modalities including: scattering off crystal planes, causing fluorescence X-ray emission from within the electron structure of the elements; and, scattering off nano-scale structures within the material being scanned.
  • the system of one of the embodiments described below provides for a detection system capable of estimating the energy of the individual X-ray photons received at the detector. This is achieved using a single detector array per X-ray source, with each of the detectors in the array constructed from an appropriate detector material coupled to a photomultiplier, producing an analog signal comprising a series of pulses - one pulse for each detected X-ray, which may or may not be overlapping when received at the detector.
  • the detector array may be arranged analogously to the freight or baggage screening systems of the prior art in order to build up an image row by row of characteristics of the item. Unlike the systems of the prior art, the detector array is capable of measuring the energy of each detected photon.
  • a pulse processing system is then used to generate a histogram for each single detector.
  • This histogram comprises of a count of the number of X-rays falling into each histogram bin in a given time interval.
  • the histogram bins represent a range of energy of the received X-rays, and the histogram is therefore the energy spectrum of the received X-ray beam.
  • the system of the described embodiments uses this full high resolution energy spectrum to obtain a much more accurate estimate of the screened material's effective atomic number (effective Z), resulting in a vastly superior classification of the screened material.
  • Figure 1 shows a high level overview of an X-Ray freight and baggage screening system according to an embodiment of the invention.
  • a host computer (103) or as shown in Figure 10 a wireless control and display system (104), for control and configuration of the X-ray screening system, and display and post processing of data collected from the X-ray scanning system.
  • a wireless control and display system for control and configuration of the X-ray screening system, and display and post processing of data collected from the X-ray scanning system.
  • Figure 2 illustrates an example diagram of the interior of the X-ray chamber, showing:
  • Figure 3 shows a more detailed view of the detection system and processing. This figure shows the steps for a single detector.
  • the effective Z may utilize, and image post processing will require, access to the spectra from all detectors.
  • a detection system and processing electronics comprising:
  • the images produced for display comprise a number of data elements recorded for each of N detector elements (501) and for each gate interval (500).
  • the data obtained for detector i during gate interval j is used in the production of effective Z, intensity and high penetration/high contrast images as shown in Figure 9 .
  • a number of elements are recorded in each pixel (502), including one or more of:
  • Figure 9 illustrates how this data is arranged and built up into an image of the scanned sample, prior to further post processing and image display.
  • the detector subsystem used in common X-ray scanning machines, for both industrial and security applications utilizes a scintillator (such as phosphor) coupled to an array of PIN diodes to convert the transmitted X-ray into light, and subsequently into an electrical signal.
  • a scintillator such as phosphor
  • detector pixels So as to achieve a resolution in the order of 1-2 mm, more than 2,000 detector pixels are used. Two separate detector arrays (and electronic readout circuits) are required for detection of the low energy X-rays and the high energy X-rays.
  • detector systems When an X-ray impacts the detector it produces an electron charge in the detector proportional to energy of the X-ray, wherein the higher the energy is the more charge is induced in the detector.
  • detector systems do not have the resolution to detect individual X-ray photons, and instead they integrate all the charge produced by the detector pixel over a given time period and convert this into a digital value. Where the instantaneous flux of X-rays on the detector pixel is large, a large digital value is produced (a bright pixel in the image) and where few X-rays impact the detector a small digital value is produced (a dark pixel in the image).
  • the detector material may be of dimensions X x Y x Z, or some other shape.
  • the photomultiplier may be a silicon photomultiplier (SiPM) and the coupling means may be a form of optical grease or optical coupling material. It may be desirable to use a form of bracket or shroud to hold the detector in position relative to the photomultiplier.
  • the photomultiplier requires appropriate power supply and bias voltage to generate the required level of amplification of the detected signal.
  • a large number of single element detector subsystems are required to produce each detector array. It may be desirable to group these in an appropriate way, depending on the specific X-Ray scanner requirements. Individual elements of detector material may be grouped into a short array of M detectors. Small groups of M detector elements may be mounted onto a single detector board, for example 2, 4 or more groups of M onto one board. The full detector array is then made up of the number of detector boards required to achieve the total number N of detector elements per array.
  • Detector subsystems can be arranged in a number of different configurations including: linear arrays of 1 x N devices; square or rectangular arrays of N x M devices; or L-shaped, staggered, herringbone or interleaved arrays.
  • a detection device used to convert incoming radiation photons into and electrical signal, is the combination of a scintillation crystal, coupled to a silicon photomultiplier (SiPM) or multi-pixel photon counter (MPPC).
  • a scintillation crystal such as LSYO (1701) is used to convert the incoming radiation photon (1700) into UV photons (1703).
  • LYSO scintillation material the peak emission of UV photons occurs at 420 nm, other scintillation material such as those listed in Table 1 may have different emission peaks.
  • a multi-pixel photon counter, or silicon photomultiplier (1704) with sensitivity in the UV region may be used to detect these photons and produce an electrical signal.
  • Figures 16A depicts a linear array of LYSO scintillation crystals (1600), indicative of how single detection devices can be joined together to form a linear array.
  • the individual LYSO crystals (1600) have a cross section of 1.8mm and a height of 5mm, the individual LYSO crystals (1600) are wrapped around the sides in a reflective material to assist in collecting all the UV photons.
  • the pitch of this exemplary array is 2.95mm, the length is 79.2mm and the width of the array is 2.5mm
  • Figures 16B and 16C depict a detector array from a top view and side view respectively, comprising the linear array of LYSO crystals depicted in 16A coupled to an electrical pulse producing element (1604) on substrate (1605).
  • the electrical pulse producing element may comprise a silicon photomultiplier (SiPM).
  • Enhanced specular reflector (ESR) or aluminium or other reflective foil (1601) is disposed around side surfaces of the scintillation crystals to direct the scintillation photons onto the silicon photomultiplier material (1604) and prevent light leakage (cross-talk) between adjacent detection devices.
  • optical coupling (1606) may be interposed between the LYSO crystals and SiPM, and may comprise any number of known suitable materials, for example, a thin layer of optically transparent adhesive.
  • scintillation crystals (1607) may be individually coupled to electrical pulse producing elements (1604), as depicted in figures 16D and 16E .
  • Coupling may be achieved by a number of methods, for example interposing an optically transparent adhesive film (1609) or optical coupling material between the scintillation crystals (1607) and electrical pulse producing elements (1604), where the electrical pulse producing elements (1604) may comprise SiPMs or an MPCC.
  • Coupling may be performed by a 'pick and place' assembly machine to individually align and couple components and coupling material.
  • Scintillation crystals may be wrapped in a reflective material such as a foil or ESR material (1608) to aid in the capture of photons.
  • the LSYO crystals (1600, 1607) may typically have a cross-section (width) approximately 1-2mm, a depth of approximately 1-2mm, and height of approximately 3-5mm, where the reflective or ESR film (1601, 1608) is approximately 0.05mm - 0.1mm thick.
  • the cross-section is 1.62mm
  • the depth is 1.24mm
  • the height is approximately 4.07mm
  • the ESR film is 0.07mm thick.
  • the cross sectional area of the scintillator material is preferably greater than 1mm square, and may be greater than 2mm square and less than 5mm square.
  • detector subsystem design uses a scintillator which is compact, robust, cost effective and non-hygroscopic, in the broadest aspect of the invention
  • detector subsystems which use alternate inorganic or inorganic scintillator materials, the characteristics of some such material are provided in Table 1.
  • Other mechanisms for converting radiation photons into electrical signals could also be considered for the detector subsystem.
  • Some examples of other detector materials options include:
  • a particular advantage of the scintillator and photomultiplier embodiment described here is the scalability of the detection elements for easy adaptability to large scanning systems such as are applicable to large freight items, which may be two or more metres in linear dimension. This is in contrast to direct conversion materials such as Cadmium Zinc Telluride, which have unacceptable dead time as the individual detector element area increases.
  • the scanning system comprises a large number of individual detectors. While each detector and associated electronics is ideally designed to have identical response to incident radiation, in practice this will not be possible. These variations between detectors result in detector to detector variation in energy spectrum output.
  • the energy spectra output from the pulse processing digital processors can be appropriately calibrated so they represent received X-ray intensity in known narrow energy bins.
  • Detector pulse calibration is used to identify the pulse characteristics for each detector required by the pulse processing system. The exact parameters required may vary, depending on the detection system.
  • the processing requires the two parameters ⁇ and ⁇ , and the pulse form p(t) which can be obtained via an appropriate calibration method, or from knowledge of the design of the detection subsystem.
  • a suitable method for estimating ⁇ , ⁇ and p(t) from received pulses is described below.
  • Each detector subsystem combined with an analog to digital converter, will have slightly different characteristics due to manufacturing variations. As a result of such component variations, the energy spectra will be scaled differently. Variations other than gain scaling are handled within the Baseline Offset Calibration or Energy Calibration.
  • the objective of the gain calibration is to achieve alignment of the energy spectra output by the pulse processing electronics across all detectors.
  • the need for absolute accuracy may be reduced or eliminated if per detector energy calibration is applied.
  • Gain calibration may be achieved in a number of ways. The following approach may be applied:
  • the detection subsystem there may be a number of detector cards, each with a number of detectors.
  • the total number of detectors may be several thousand or more. Results from one example of such a detector board are presented here.
  • the example board comprises 108 detectors, with in this case LYSO used as the scintillator material. These detectors are packed into linear arrays of 27 detectors. Each detector board then uses 4 x 27 detector arrays to achieve a total of 108 detectors.
  • each detector When X-Rays are incident upon a detector, photons are emitted by the LYSO based on the energy of the incident X-Ray.
  • Each detector is placed above a SiPM, and it is the SiPM that detects and amplifies the emitted photons.
  • the detectors are coupled to the SiPM via an optical grease.
  • the gain of each SiPM is determined by the bias voltage applied, and the SiPM breakdown voltage.
  • Each detector subsystem may have a slightly different baseline level, as measured at the output of the Analog to digital converter.
  • the baseline is estimated and removed. Any suitable method can be used including, for example:
  • the pulse processing electronics will produce an energy spectrum that is uncalibrated. That is, the output will comprises a number of counts in a set of histogram bins, but the exact energy of those histogram bins is unknown. In order to achieve accurate effective Z results, knowledge of the energy of each bin is required.
  • individual calibration table / calibration parameters can be generated for each detector.
  • any suitable method can be used for this calibration, including injecting a known source spectrum of variable count rate, and recording the spectrum shift as count rate increases.
  • the source has a narrow energy band so the shift can be clearly measured, and also variable energy so the offset can be calibrated as a function of energy if desirable.
  • the residual spectrum is measured with a large mass of material in the beam, sufficient to completely block the X-ray beam, such as a large thickness of steel. In practice, a small level of energy still reaches the detector array, whether from scatter or other mechanisms, and this residual spectrum must be measured so it can be removed from the received spectra during normal operation.
  • the residual spectrum is then measured by averaging the received spectra for a number of gate intervals with the blocking mass in the beam.
  • the pileup parameters can be calibrated in several ways, for example:
  • a high rate pulse processing system (305), such as those disclosed in US Patent No. 7383142 , US Patent No. 8812268 or WO/2015/085372 , is allocated to each detector subsystem, to perform the following operations on the digitized pulse signal output from the analog to digital converter:
  • the thresholds R low and R high may be pre-set or user configurable.
  • intensity images with varying contrast are generated based on integrating the received spectrum across different energy bands.
  • the system can only utilize the broad energy range inherent in the detector material.
  • arbitrary energy ranges can be used to generate associated intensity images in that energy range.
  • Specific energy ranges can then be defined in order to best isolate and display particular material types, with energy ranges tuned, for example, for organic material, inorganic material, or light, medium or heavy metals.
  • the energy band may be user defined or pre-configured. One, two or more different energy bands may be configured to enable the user to select between images of interest.
  • the effective Z processing involves the use of full energy spectra computed by the pulse processing electronics, combined with the energy calibration, to compute an estimate of the effective Z of the sample material.
  • the effective Z processing is performed for every detector, and for each detector proceeds as follows (so for a 1 x N detector array, this process is repeated N times). To reduce computational requirement, the effective Z processing is only performed for received detectors i and gate intervals j that are not declared either impenetrable or empty.
  • the accumulation of reference spectra ceases and the average of M collected spectra can be used for the reference, or the measurement terminated if M is insufficient.
  • the mass attenuation constants for a given effective Z and given energy define the extent to which the given material Z will attenuate X-rays of energy E.
  • Mass attenuation data is available at a finite (small) number of energies, perhaps every 10, 20 or 50 keV, whereas the energy spectra created by the method disclosed in this disclosure may be generated at energy spacing as little as 1 keV or even less. In practice a finite number of these energy values will be selected for use in the effective Z computation.
  • Z there is no particular requirement for effective Z to be integer, and in fact the mass attenuation table may contain values for non-integer values of Z representing composite materials. However, it is clearly not possible to represent a continuum of possible Z values in a finite table. In order to compute Z to arbitrary precision, it is possible to interpolate the cost function to the required resolution using an appropriate interpolation algorithm. The value of Z chosen is then the value which minimizes the interpolated cost function.
  • the cost function C(Z) is a smooth function, and therefore an actual floating point or continuous value of Z which minimises this smooth function can be reliably predicted from the curve via some form of interpolation.
  • step 3 above indicates the cost function is computed for all available Z values in the mass attenuation table.
  • efficient search methods can be applied to reduce the computational requirements. Such methods include one or more of the following:
  • the cost function form has been chosen so as to be relatively insensitive to noise on the spectrum.
  • an alternative method has been developed whereby the system is calibrated using varying thickness samples of known materials.
  • the aim is to calibrate the expected received spectra as a function of material, material thickness, and energy histogram bins. This avoids the requirement for absolute energy calibration, and also largely avoids the effect of spectrum shift with count rate (if present). The need for pileup removal may also be eliminated.
  • the received spectra from all detectors are consistent with each other, and so it is only desirable to obtain calibration data at one detector for use at all detectors. In practice, it is likely to be desirable to obtain calibration data for groups of adjacent detectors or possibly every detector, depending on the consistency between detectors.
  • the first step in the calibration process is to obtain a reference spectrum I 0 (B) at each histogram bin B, with no material in the X-Ray beam for the detector(s) to be calibrated. Histogram bins will now be denoted by B rather than E to denote that there is no requirement to calibrate the bins in terms of their exact energy.
  • the tables of Tx(Z, B, x) and R(Z, x) together form the calibration tables that are used to estimate effective Z at each pixel (detector/gate interval). As previously stated, they may or may not be a function of detector also, depending on the equivalence of data from all detectors.
  • Z there is no particular requirement for effective Z to be integer, and in fact the self calibration table may contain values for non-integer values of Z representing composite materials. However, it is clearly not possible to represent a continuum of possible Z values in a finite table. In order to compute Z to arbitrary precision, it is possible to interpolate the cost function to the required resolution using an appropriate interpolation algorithm. The value of Z chosen is then the value which minimizes the interpolated cost function.
  • the cost function C(Z) is a smooth function, and therefore an actual floating point or continuous value of Z which minimises this smooth function can be reliably predicted from the curve via some form of interpolation.
  • the cost function C(Z) is a smooth function, and therefore an actual floating point or continuous value of Z which minimises this smooth function can be predicted from the curve via some form of interpolation.
  • the quadratic model is just a model to ensure a consistent effective Z is obtained for a particular material. It is not intended to be an accurate functional model of the cost function behaviour, and it is not considered necessary.
  • the principle objective is to obtain an estimate of effective Z that is consistent for a particular material, and enables reliable separation of closely spaced materials.
  • the quadratic model achieves this objective.
  • the floating point effective Z algorithm was tested on a range of material samples and also tested extensively on the briefcase. There were several observations made about the performance.
  • Figure 20 indicates an overview of the various optional processing stages that may be implemented in the present method.
  • the tiling algorithm is effectively a block averaging function.
  • the purpose of the tiling algorithm is to average the floating effective Z image over an area (mm 2 ) that represents the smallest object required to be detected of a constant intensity and material composition.
  • the tiling algorithm generates tiles with 50% overlap to ensure that we always capture the object of interest.
  • the tiling algorithm estimates the mean and standard deviation over rectangular tiles in the floating effective Z image.
  • the tile width and height are defined by the user. Tiles are overlapped by 50% in both vertical and horizontal dimensions. Given an image size Nr by Nc pixels, and a tile dimension Tr by Tc pixels, the number of tiles in the vertical dimension is floor (Nr/Tr) ⁇ 2. The tile dimensions must be even valued to ensure 50% overlap.
  • the tiling algorithm executes a loop that indexes into each tile and calculates the mean and standard deviation of all pixels in the tile.
  • the clustering algorithm groups tiles that have a common effective Z and are spatially connected.
  • the purpose of clustering algorithm is to detect objects that span areas larger than the minimum object size as defined by the tile dimensions, see section 2.1. Connectedness is defined along edges. Connected tiles are assigned a common cluster ID.
  • the output of the clustering algorithm is a cluster map and a cluster table.
  • the cluster map is a matrix of connected tiles with associated cluster IDs.
  • the cluster table holds information on each cluster ID including the number of tiles in the cluster, and the vertical and horizontal extent of each cluster.
  • the sets A and B are adapted at eight boundary conditions, four along the image edges and four at the image vertices.
  • Figure 19 depicts the formation of clusters, where single tiles are ignored.
  • the threat detection algorithm is a nearest neighbour classifier.
  • the algorithm classifies individual tiles. There are two steps in the algorithm, training and classification.
  • the training stage establishes a lookup table mapping normalized intensity to floating effective Z for a range of materials that are referred to as 'threats'. This terminology is of no consequence.
  • the lookup table simply contains materials of interest. In the current implementation, the lookup table is approximated as a quadratic fit, for which only the quadratic coefficients are stored (see threat.cpp).
  • the input is the normalized measured tile intensity (Imeas), the measured tile floating effective Z (Zmeas), and a maximum effective Z classification error (deltaZ).
  • the classifier declares positive classification if abs(C i (Imeas)-Zmeas) ⁇ deltaZ, where C i is the quadratic function associated with the i th threat material.
  • the use of both intensity and effective Z in the threat profile is an important aspect of this approach.
  • the effective Z is typically not constant with material thickness, and so including the intensity (related to thickness) provides a two dimensional test with far superior discrimination than effective Z alone.
  • Figure 15 shows the effective Z vs intensity for a range of material samples tested, along with the quadratic interpolation. Here the variation of effective Z with intensity is clear.
  • the purpose of the edge detection algorithm is to ensure that the moving average window in section 2.5 does not straddle material boundaries.
  • the edge detection uses amplitude transitions in the intensity image to declare material edges.
  • the input to the edge detection algorithm is the intensity image. Edges are only detected in the horizontal dimension. The reason for not detecting edges in the vertical dimension is that the moving average window only operates in the horizontal dimension. Edges in the intensity image are computed for each detector.
  • a first order gradient operator is used to detect edges.
  • An edge is declared when abs(G) > g where g is a user defined threshold.
  • the purpose of the moving average algorithm is to filter the intensity histograms for each detector so as to increase the effective signal-to-noise ratio.
  • the algorithm generates a filtered intensity histogram a slice k, for each detector, by averaging the measured intensity histograms over a symmetric window centred on slice k.
  • the edge detector plays an important role in ensuring the moving average window does not straddle different materials. If a window overlaps an edge the average is only calculated up to the edge boundaries. The width of the window can be set by the user. On edges, no averaging is performed.
  • Figure 22 illustrates the behaviour of the moving average as it transitions over an edge.
  • a source of Gamma-rays (1800), such as Cobalt 60 may be used to irradiate the tunnel of a scanner (1801) with Gamma-ray photons.
  • the Gamma-rays source (1800) may be shielded (1802) and a collimator (1803) may also be used to create a fan beam of Gamma-rays (1804).
  • a system of rollers (1805) or other devices such as conveyors may be use to pass cargo (1806), parcels, bags or other items of interest through the fan beam of Gamma-rays (1804).
  • the Gamma-ray photons will interact with the cargo (1806) via a range of interactions including absorption, scattering and recoil.
  • Gamma-ray photons which pass through the cargo may be detected on the other side of the scanner by a detector subsystem.
  • a detector subsystem (1807) may be an array of scintillation detectors coupled to silicon photomultipliers to produce and electrical signal.
  • the array may consist of semiconductor material such as High Purity Germanium (HPGe), which is capable of direct conversion of Gamma-ray photons into an electrical charge.
  • HPGe High Purity Germanium
  • any suitable method of high rate pulse processing can be used within the embodiments described herein.
  • the high X-Ray flux present in typical X-Ray screening systems results in a high pulse count rate, and a high likelihood of receiving overlapping X-Ray pulses.
  • Pulse pile-up has long been a problem to contend within applications of high rate radiation spectroscopy.
  • Traditional approaches to pulse shaping use linear filters to shorten pulse duration which can significantly reduce SNR and are therefore limited to output rates of a few hundred kc/s.
  • An alternate approach to processing the data from radiation detectors is based on the idea of mathematically modeling data corrupted by pulse pile-up and solving for the required model parameters. By recovering rather than discarding data corrupted by pulse pile-up this technique enables high throughput, low dead-time pulse processing without the traditional loss in energy resolution.
  • the digitized radiation detector time series ( y ) is modeled as the sum of an unknown number of radiation events ( N ), with random times of arrival ( ⁇ ), and amplitudes (a), interacting with a radiation detector, that have an expected pulse shape (h) and with a noise process ( ⁇ ).
  • the digitized detector data can be accurately decomposed into the individual component events and the energy of each event determined.
  • Calibration of the detector is the first stage of the algorithm; it takes as input the detector time series data and determines the unit impulse response of the detector (the expected pulse shape from the detector). Refer to Pulse Parameter Calibration for a more detailed summary of the pulse calibration process.
  • the Pulse Localization stage After the unit impulse response of the detector has been determined this is used by the Pulse Localization stage to determine the number of events in the digitized detector data stream and their TOA relative to each other.
  • the detection of events in the digitized detector waveform is accomplished by fitting an exponential model to a fixed number of data points. After the System Characterization stage the exponential decay of the pulse tail is well characterized.
  • the detection metric (the signal ultimately used to make a decision as to whether a pulse has arrived or not) is formed by fitting an exponential curve to a specified number of data points. This fixed length 'detection window' is run continuously over the digitized detector data and the sum of the squares of the error is computed (this can also be thought of as the sum of the square of the fit residual). This operation results in three distinct modes of operation:
  • the final stage of Pulse Localization is making a decision on the exact number and time of arrival of each of the radiation events in the detector data stream.
  • One approach would be to apply a simple threshold to the detection metric and declare a pulse arrival at the nearest sample to the threshold crossing.
  • a simple threshold crossing is susceptible to noise and only provides ⁇ 0.5 sample accuracy in determining the pulse arrival time.
  • a quadratic peak detection algorithm can be used. Such an approach fits a quadratic to a sliding window of N samples of the detection metric (N maybe equal to 5).
  • the Pulse Energy Estimation stage determines the energy of all the radiation events in the detector data stream. As its input it uses: the a priori knowledge of the detector unit impulse response; the number of events; and their individual time of arrival data.
  • the columns of matrix A contain multiple versions of the unit impulse response of the detector.
  • the starting point of the signal shape is defined by the signal temporal position. For example, if the signals in the data arrive at positions 2, 40, 78 and 125, column 1 of matrix A will have '0' in the first row, the 1 st data point of the unit impulse response in the second row, the 2 nd data point of the unit impulse response in the 3 rd row, etc.
  • the second column will have '0' up to row 39 followed by the signal form.
  • the third column will have '0' up to row 77; the fourth column will have '0' up to row 124 and then the signal form.
  • the size of matrix A is determined by the number of identified signals (which becomes the number of columns), while the number of rows depends on the number of samples in the 'time series'.
  • the final functional stage of the real-time, signal-processing algorithm is the Validation stage. At this stage all the parameters that have been estimated by previous algorithmic stages (pulse shape, number of events, time of arrival and event energy) are combined to reconstruct a 'noise-free' model of the detector data.
  • the accuracy of the estimated parameters can be determined. Much like examining the residual from a straight line fit of a data set, if the magnitude of the residuals is small, the parameters well describe the data. However, if large residuals are observed, the detector data has been poorly estimated and that portion of the data can be rejected.
  • the method for resolving individual signals in detector output data comprises:
  • the signal form may generally be regarded as characterising the interaction between the detector and the radiation (or other detected input) that was or is being used to collect the data. It may be determined or, if known from earlier measurements, calibrations or the like, obtained from (for example) a database.
  • transforming the digital series according to the mathematical transform comprises forming a model of the digital series and transforming the model of the digital series according to the mathematical transform.
  • the method includes determining a plurality of parameters of the transformed signals, such as frequency and amplitude.
  • the transform is a Fourier transform, such as a fast fourier transform or a discrete fourier transform, or a wavelet transform.
  • the transform may be applied somewhat differently to the signal form and digital series respectively.
  • the mathematical transform is the Fourier transform, but the signal form is transformed with a discrete fourier transform and the digital series is transformed with a fast fourier transform.
  • 'corrupt data' some data (which hence is termed 'corrupt data'), as is described below.
  • 'signal' is interchangeable in this context with 'pulse', as it refers to the output corresponding to individual detection events rather than the overall output signal comprising the sum of individual signals.
  • the temporal position (or timing) of a signal can be measured or expressed in various ways, such as according to the time (or position in the time axis) of the maximum of the signal or the leading edge of the signal. Typically this is described as the arrival time ('time of arrival') or detection time.
  • 'detector data' refers to data that has originated from a detector, whether processed subsequently by associated or other electronics within or outside the detector.
  • the signal form may be determined by a calibration process that involves measuring the detector's impulse response (such as time domain response or frequency domain response) to one or more single event detections to derive from that data the signal form or impulse response.
  • a functional form of this signal form may then be obtained by interpolating the data with (or fitting to the data) a suitable function such as a polynomial, exponential or spline.
  • a filter (such as an inverse filter) may then be constructed from this detector signal form.
  • An initial estimate of signal parameters may be made by convolution of the output data from the detector with the filter.
  • Signal parameters of particular interest include the number of signals and the temporal position (or time of arrival) of each of the signals.
  • the particular signal parameters of interest can then be further refined.
  • the accuracy of the parameter estimation can be determined or 'validated' by comparing a model of the detector data stream (constructed from the signal parameters and knowledge of the detector impulse response) and the actual detector output. Should this validation process determine that some parameters are insufficiently accurate, these parameters are discarded. In spectroscopic analysis using this method, the energy parameters deemed sufficiently accurate may be represented as a histogram.
  • the data may include signals of different forms.
  • the method may include determining where possible the signal form of each of the signals.
  • the method includes progressively subtracting from the data those signals that acceptably conform to successive signal forms of a plurality of signal forms, and rejecting those signals that do not acceptably conform to any of the plurality of signal forms.
  • the method may further comprise detecting a pulse or pulses in said detector output data by:
  • the one or more functions are functions of time.
  • the one or more functions may not be functions exclusively of time.
  • the method may comprise providing the detector output data in, or converting the detector output data into, digital form before fitting the one or more functions to the detector output data.
  • v t 1 1 ⁇ e ⁇ ⁇ ⁇ ⁇ ⁇ e ⁇ ⁇ t ⁇ e ⁇ ⁇ t whenever ⁇ ⁇ ⁇
  • Solving equation (1) can be done numerically, such as with a bisection method, especially since the left hand side is monotonic in ⁇ . Determining the left hand side for different values of ⁇ may be done by any suitable technique, such as with a Taylor series expansion for small ⁇ . (In practice, the value of ⁇ will generally be small because noise will generally preclude accurate characterization of a pulse that started in the distant past.)
  • the method may comprise constraining ⁇ by requiring that ⁇ ⁇ [ ⁇ *, 0].
  • ⁇ ⁇ [ ⁇ *, 0] is equivalent to the constraint on a and b that 0 ⁇ b ⁇ ca
  • the method may also comprise constraining the amplitude of the pulse. This can be used, for example, to prevent a fitted pulse from being too small or too large. Indeed, referring to equation (2) above, if ⁇ is constrained to lie between -1 and 0 then A lies between ⁇ -1 a and ⁇ -a e ⁇ a. Constraining a therefore constrains the amplitude A .
  • the function f is in the form of a function with three exponentials.
  • the time constants ⁇ 1 , ⁇ , ⁇ 3 are known and dissimilar (so fewer problems of numerical imprecision arise), and the method includes fitting the curve: A 1 e ⁇ ⁇ 1 t + ⁇ + A 3 e ⁇ ⁇ 3 t .
  • the time constants ⁇ 1 , ⁇ , ⁇ 3 are known and in ascending order such that ⁇ 1 ⁇ ⁇ 2 ⁇ ⁇ 3
  • v 1 t ⁇ 31 ⁇ ⁇ 21 ⁇ 32 1 ⁇ 31 ⁇ 21 e ⁇ ⁇ 1 t ⁇ 1 ⁇ 32 ⁇ 21 e ⁇ ⁇ 2 t + 1 ⁇ 32 ⁇ 31 e ⁇ ⁇ 3 t
  • v 2 t 1 ⁇ 21 e ⁇ ⁇ 1 t ⁇ e ⁇ ⁇ 2 t
  • v 3 t e ⁇ ⁇ 1 t .
  • the function f may be a superposition of a plurality of functions.
  • the present invention relates generally to a method and apparatus for estimating the location and amplitude of a sum of pulses from noisy observations of detector output data. It presented the maximum-likelihood estimate as the benchmark (which is equivalent to the minimum mean-squared error estimate since the noise is additive white Gaussian noise).
  • the method may comprise low-pass filtering the data before fitting the one or more functions.
  • the method comprises forcing any estimates having the pulse starting within the window to start at a boundary of the window.
  • the method comprises maximizing window size or varying window size.
  • the method comprises transforming the detector output data with a transform before fitting the one or more functions to the detector output data as transformed.
  • the method may also comprise subsequently applying an inverse transform to the one or more functions, though in some cases it may be possible to obtain the desired information in the transform space.
  • the transform may be a Laplace transform, a Fourier transform or other transform.
  • estimating the location of the peak comprises minimizing an offset between the start of a window and a start of the pulse.
  • the method further comprises detecting a pulse or pulses in the data by:
  • the invention provides a method for locating a pulse in detector output data, comprising:
  • each of the one or more functions is a superposition of a plurality of functions.
  • resolving individual signals in the detector output data comprises transforming detector data to produce stepped data, or using data that is already in a stepped form, and detecting at least one signal and estimating a parameter of the signal based at least partially on the stepped data.
  • the method comprises transforming the detector output data to produce stepped data or integral data, detecting at least one event, and estimating a pulse energy associated with the event.
  • detecting the at least one event occurs by fitting an expected pulse shape with a sliding window segment of the transformed pulse shape data.
  • the method further comprises the step of detecting peaks in the signal, wherein a detection metric is applied to the transformed data.
  • the detection metric is compared to a simple threshold - if the metric is less than the threshold, then no pulses are deemed present - if it exceeds the threshold, then one or more pulses may be present.
  • Declaration of significant peaks in the detection metric is conducted, when the slope of the peak changing from positive to negative indicates an event.
  • 'corrupt data' uncharacterized data
  • 'pulse' the term 'signal' is interchangeable in this context with 'pulse', as it refers to the output corresponding to individual detection events rather than the overall output signal comprising the sum of individual signals.
  • the temporal position (or timing) of a signal can be measured or expressed in various ways, such as according to the time (or position in the time axis) of the maximum of the signal or the leading edge of the signal. Typically this is described as the arrival time ('time of arrival') or detection time.
  • 'detector data' refers to data that has originated from a detector, whether processed subsequently by associated or other electronics within or outside the detector.
  • the method optionally comprises deleting samples within a set window around the rising edge to ensure the edge region of each pulse, where the real transformed pulse data differs from an ideally transformed pulse, is excluded from the calculations.
  • the method optionally comprises an assessment of variance of the energy estimations in the data, and validation of the modeled data.
  • the method may include building a model of the data from the processed data output and determining the accuracy of the modeling based on a comparison between the detector output data and the model.
  • the method includes creating a model of the detector output using the signal parameters in combination with the detector impulse response.
  • the method may include performing error detection by comparing the actual detector output with the model of the detector output, such as by using least squares.
  • the method may include discarding energy estimates deemed not sufficiently accurate.
  • the method includes presenting all sufficiently accurate energy estimates in a histogram.
  • the majority of the FFT bins are discarded, as it is only necessary to keep approximately 1/8 of the FFT bins in order to accurately reconstruct the energy spectrum. For example if there are 512 histogram bins computed, only 32 complex FFT bins are retained. The last 32 complex FFT bins are just the complex conjugate of these bins, and the remaining 448 bins contain (almost) no information.
  • Figure 8 illustrates the result achieved.
  • the rectangular window if used on its own, results in the measured spectrum being convolved with a sinc function, with width at the mid amplitude of approx 10 samples, but significant oscillations - around 22% at the first negative going peak.
  • user window w it is possible to achieve a "time" domain response that is approximately raised cosine in nature, with very little oscillatory nature - around 0.2%.
  • the width at mid amplitude increases to around 20 samples.
  • the additional benefit of an appropriately designed window function is that the received energy spectra are smoothed before processing, resulting in a significant reduction in noise in the effective Z estimates, and the potential for using less bins in the effective Z estimate while achieving a similar result.
  • p t ⁇ t ⁇ T a t A exp ( ⁇ ⁇ ⁇ ⁇ t 0 ⁇ exp ( ⁇ ⁇ ⁇ ⁇ t 0 d ⁇
  • a and ⁇ are the falling edge and rising edge time constants respectively
  • t 0 is the pulse time of arrival
  • T a is the pulse averaging window
  • A is a pulse scaling factor related to the pulse energy.
  • the pulse parameters may be estimated from a time series capture of the digitized detector signal as follows:
  • DC offset or signal baseline, used interchangeably
  • This DC offset can arise from various sources including the bias levels of analog electronics, the analog to digital conversion, and the detector itself.
  • Control theory suggests the DC offset error may be tracked and reduced to zero by generating a feedback signal that is proportional to the integral of the signal - however there is a significant problem in the case of pulse processing. Pulses introduce additional features to the signal that have non-zero mean. This introduces a bias dependent on pulse energy, count rate and pulse shape, which corrupts the feedback signal and prevents standard control loop tracking from successfully removing the DC offset.
  • the detector signal output is digitally processed to remove the pulse shaping effects introduced by analog electronics.
  • this processed signal results in a signal shape that has constant value in the regions between pulse arrivals, and a rapid change in value where pulses arrive. If a residual DC offset is present in the detector signal the processed signal changes linearly with time in the regions between pulse arrivals.
  • An error feedback signal that is proportional to the slope of this signal may be formed by taking the difference between two samples. These samples need not be consecutive, but may be separated by 'N' samples in time. By choosing an appropriate value for 'N', a signal with a suitable signal to noise ratio may be found for driving a feedback look.
  • the baseline tracking loop is not updated when a pulse has arrived between the two samples used to generate the feedback error signal.
  • bias may be further reduced by preventing the baseline tracking loop from updating when a pulse has arrived within a guard region on either side of the samples used to generate the feedback error signal.
  • the value of the processed detector signal increases whenever a pulse arrives. This eventually causes the internal registers used to store the value of the signal to overflow. The value of the processed signal is monitored, and when overflow is detected, the baseline tracking loop is prevented from updating until the effects of overflow have passed.
  • the same hardware can be used for tracking multiple baseline offsets in multiple channels in a time division multiplexed scheme.
  • the values for the tracking loop variables for each channel are stored/loaded when switching between channels.
  • the baseline tracking loop is prevented from updating until transient effects of the detector channel change has passed.
  • Very tight collimation may be used within the scanner in order to minimise the effect of scatter on the measured spectrum. This is particularly important where transitions from high to low or low to high intensity occur.
  • the overall results of the system have shown that scatter has been largely addressed through the inclusion of tight collimation.
  • the purpose of the reference calculation is to establish the mean intensity for each detector. This value is used to scale all intensity histograms to unit energy. This is commonly referred to as normalization.
  • a reference intensity is calculated for each detector. The reference intensity is calculated as the mean intensity over the first N slices in a scan. The intensity is the 1st bin in the FFT or the sum of all complex-valued elements in the FFT vector.
  • the reference histogram is also a reference histogram computed in the same way - by averaging the measured energy histograms for the first N slices.
  • the reference histogram is used to normalise all measured histograms to ensure any run to run variations in X-ray flux do not impact the effective Z computation.
  • the reference is measured during an interval where:
  • the current implementation uses a duration measured in slices. This can cause problems when the slice rate is slowed below 5 ms for example - the reference collection can run into the sample under test. This needs to be corrected in 2 ways to be fully robust:
  • More accurate and consistent effective Z may be obtained if a longer reference collection duration was used.
EP15862380.1A 2014-11-30 2015-11-28 Device and method for material characterisation Active EP3224603B1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
AU2014268284A AU2014268284A1 (en) 2014-11-30 2014-11-30 Method and apparatus for material identification
PCT/AU2015/050752 WO2016082006A1 (en) 2014-11-30 2015-11-28 Device and method for material characterisation

Publications (3)

Publication Number Publication Date
EP3224603A1 EP3224603A1 (en) 2017-10-04
EP3224603A4 EP3224603A4 (en) 2018-07-25
EP3224603B1 true EP3224603B1 (en) 2021-04-28

Family

ID=56073233

Family Applications (1)

Application Number Title Priority Date Filing Date
EP15862380.1A Active EP3224603B1 (en) 2014-11-30 2015-11-28 Device and method for material characterisation

Country Status (12)

Country Link
US (2) US10416342B2 (zh)
EP (1) EP3224603B1 (zh)
JP (1) JP6875277B2 (zh)
CN (1) CN107250777B (zh)
AU (2) AU2014268284A1 (zh)
CA (1) CA2968883A1 (zh)
DK (1) DK3224603T3 (zh)
EA (1) EA036838B1 (zh)
ES (1) ES2877564T3 (zh)
HK (1) HK1244535A1 (zh)
IL (1) IL251995B (zh)
WO (1) WO2016082006A1 (zh)

Families Citing this family (28)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6548565B2 (ja) * 2015-12-14 2019-07-24 浜松ホトニクス株式会社 シンチレータパネル、及び、放射線検出器
JP2017130334A (ja) * 2016-01-20 2017-07-27 株式会社日立ハイテクノロジーズ 荷電粒子ビーム装置及び荷電粒子ビーム装置の画像形成方法
WO2017205914A1 (en) * 2016-05-30 2017-12-07 Southern Innovation International Pty Ltd Material characterisation system and method
EP3465177A4 (en) 2016-05-30 2020-03-04 Southern Innovation International Pty Ltd MATERIAL CHARACTERIZATION SYSTEM AND METHOD
WO2018081166A1 (en) * 2016-10-24 2018-05-03 Xia Llc Ratio-reference measurement of the arrival time and/or amplitude of a digitized electronic pulse
US10817588B2 (en) 2016-10-24 2020-10-27 Xia Llc Ratio-reference measurement of the arrival time and/or amplitude of a digitized electronic pulse
US10416295B2 (en) 2016-10-24 2019-09-17 Xia Llc Interpolation measurement of the arrival time and/or amplitude of a digitized electronic pulse
EP3548877A4 (en) * 2016-11-29 2020-08-12 Laitram, L.L.C. MULTI-ENERGY X-RAY ABSORPTION IMAGING FOR DETECTING FOREIGN BODIES ON A CONVEYOR
US11120104B2 (en) * 2017-03-01 2021-09-14 Stmicroelectronics (Research & Development) Limited Method and apparatus for processing a histogram output from a detector sensor
DE102017003517A1 (de) 2017-04-11 2018-10-11 Universität Hamburg Verfahren und Messvorrichtung zur Röntgenfluoreszenz-Messung
AU2018202912B1 (en) * 2018-04-27 2019-06-20 Southern Innovation International Pty Ltd Input count rate estimation in radiation pulse detectors
GB201807616D0 (en) * 2018-05-10 2018-06-27 Radio Physics Solutions Ltd Improvements in or relating to threat classification
WO2020068178A2 (en) * 2018-05-18 2020-04-02 Lawrence Livermore National Security, Llc Multifaceted radiation detection and classification system
GB2577686B (en) * 2018-10-01 2022-08-10 Smiths Heimann Sas Correction of images
US10908101B2 (en) * 2018-11-16 2021-02-02 Core Laboratories Lp System and method for analyzing subsurface core samples
JP7242266B2 (ja) * 2018-11-29 2023-03-20 キヤノン株式会社 放射線撮像装置および放射線撮像装置の制御方法
WO2020160106A1 (en) * 2019-01-30 2020-08-06 Ndc Technologies Inc. Radiation-based thickness gauge
JP7309385B2 (ja) 2019-02-28 2023-07-18 富士フイルムヘルスケア株式会社 光子計数型検出器の較正方法
DE102019111567A1 (de) * 2019-05-03 2020-11-05 Wipotec Gmbh Verfahren und Vorrichtung zur Röntgeninspektion von Produkten, insbesondere von Lebensmitteln
CN110431407B (zh) 2019-06-20 2020-08-25 长江存储科技有限责任公司 多晶硅表征方法
CN110849971B (zh) * 2019-11-21 2021-05-18 西南交通大学 基于双指数窗函数法的结构模态参数识别方法
CN111708071B (zh) * 2020-06-05 2022-08-05 成都理工大学 核废物包装体流水线式扫描检测装置
US11879854B2 (en) * 2020-09-23 2024-01-23 Baker Hughes Oilfield Operations Llc Positioning of x-ray imaging system using an optical camera
JP7178725B2 (ja) * 2020-11-30 2022-11-28 株式会社リガク 蛍光ｘ線分析装置
JP2022094375A (ja) 2020-12-15 2022-06-27 株式会社日立ハイテク 荷電粒子線装置、計算機および荷電粒子線装置の信号処理方法
CN113310442B (zh) * 2021-04-27 2023-04-28 长江存储科技有限责任公司 一种厚度的测量方法及其装置
CN114186596B (zh) * 2022-02-17 2022-04-22 天津国科医工科技发展有限公司 一种谱图峰的多窗口识别方法、装置以及电子设备
US20230147681A1 (en) * 2022-07-26 2023-05-11 Rapiscan Holdings, Inc. Methods and Systems for Performing On-The-Fly Automatic Calibration Adjustments of X-Ray Inspection Systems

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100034353A1 (en) * 2008-08-11 2010-02-11 Kravis Scott D Scanning X-ray inspection system using scintillation detection with simultaneous counting and integrating modes

Family Cites Families (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE59308726D1 (de) * 1992-02-06 1998-08-13 Philips Patentverwaltung Anordnung zum Messen des Impulsübertragsspektrums von elastisch gestreuten Röntgenquanten
JPH0933661A (ja) * 1995-07-25 1997-02-07 Osaka Gas Co Ltd 相関型探知方法及び装置
US5600700A (en) * 1995-09-25 1997-02-04 Vivid Technologies, Inc. Detecting explosives or other contraband by employing transmitted and scattered X-rays
DE60135568D1 (de) 2000-10-11 2008-10-09 Symetrica Ltd Gammastrahlenspektrometrie
EP1579202B1 (en) * 2002-12-10 2007-04-18 Commonwealth Scientific And Industrial Research Organisation Radiographic equipment
AU2002953244A0 (en) * 2002-12-10 2003-01-02 Commonwealth Scientific And Industrial Research Organisation A detection system
EP1733254A4 (en) * 2004-03-01 2011-05-18 Varian Med Sys Inc EXAMINATION OF OBJECTS BY SCANNING USING DOUBLE-ENERGY RAYS AND DETECTION OF DELAYED NEUTRONS
JP5022902B2 (ja) 2004-09-16 2012-09-12 サザン イノヴェーション インターナショナル プロプライアトリー リミテッド 検出器出力データ内の個別信号分離装置および方法
US7769132B1 (en) * 2007-03-13 2010-08-03 L-3 Communications Security And Detection Systems, Inc. Material analysis based on imaging effective atomic numbers
JP5832892B2 (ja) 2008-03-31 2015-12-16 サザン イノヴェーション インターナショナル プロプライアトリー リミテッド 孔内検層のための方法および装置
AU2009230874B2 (en) 2008-03-31 2014-06-26 Southern Innovation International Pty Ltd Screening method and apparatus
AU2009230876B2 (en) 2008-03-31 2014-07-10 Southern Innovation International Pty Ltd Radiation imaging method with individual signal resolution
US8812268B2 (en) 2008-12-18 2014-08-19 Southern Innovation International Pty. Ltd. Method and apparatus for resolving piled-up pulses by using a mathematical transform
CN102884422B (zh) * 2010-02-25 2016-09-28 拉皮斯坎系统股份有限公司 用以确定物质的原子序数的、高能量x射线的基于光谱学的检查系统和方法
US9224573B2 (en) * 2011-06-09 2015-12-29 Rapiscan Systems, Inc. System and method for X-ray source weight reduction
US9218933B2 (en) * 2011-06-09 2015-12-22 Rapidscan Systems, Inc. Low-dose radiographic imaging system
US9922003B2 (en) * 2011-06-14 2018-03-20 Southern Innovation International Pty Ltd Method and apparatus for identifying pulses in detector output data
CN103135123A (zh) * 2011-11-30 2013-06-05 中国辐射防护研究院 基于硅光电倍增器的环境X、γ辐射测量方法及测量装置
US8907290B2 (en) * 2012-06-08 2014-12-09 General Electric Company Methods and systems for gain calibration of gamma ray detectors
JP6111506B2 (ja) * 2013-05-07 2017-04-12 パナソニックＩｐマネジメント株式会社 センサ装置
MX364668B (es) 2013-09-19 2019-05-03 Rapiscan Systems Inc Sistema de inspeccion radiografica de dosis baja.
CA2933255C (en) 2013-12-11 2023-08-15 Southern Innovation International Pty Ltd Method and apparatus for resolving signals in data

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100034353A1 (en) * 2008-08-11 2010-02-11 Kravis Scott D Scanning X-ray inspection system using scintillation detection with simultaneous counting and integrating modes

Also Published As

Publication number Publication date
HK1244535A1 (zh) 2018-08-10
AU2015354412B2 (en) 2021-07-01
JP2018502288A (ja) 2018-01-25
WO2016082006A1 (en) 2016-06-02
EA036838B1 (ru) 2020-12-25
EP3224603A4 (en) 2018-07-25
ES2877564T3 (es) 2021-11-17
US10545257B2 (en) 2020-01-28
JP6875277B2 (ja) 2021-05-19
EA201791200A1 (ru) 2017-09-29
US20170269257A1 (en) 2017-09-21
IL251995B (en) 2021-12-01
CN107250777A (zh) 2017-10-13
US20190346586A1 (en) 2019-11-14
EP3224603A1 (en) 2017-10-04
CN107250777B (zh) 2020-10-23
US10416342B2 (en) 2019-09-17
AU2015354412A1 (en) 2017-05-04
DK3224603T3 (da) 2021-07-19
CA2968883A1 (en) 2016-06-02
AU2014268284A1 (en) 2016-06-16
IL251995A0 (en) 2017-06-29

Similar Documents

Publication Publication Date Title
US10545257B2 (en) Device and method for material characterization
US11243327B2 (en) System and method for material characterization
AU2014361753B2 (en) Method and apparatus for resolving signals in data
US11796479B2 (en) System and method for material characterization
Mohammadian-Behbahani et al. Pile-up correction algorithm based on successive integration for high count rate medical imaging and radiation spectroscopy
Meng et al. Exploring the limiting timing resolution for large volume CZT detectors with waveform analysis
CN106124539B (zh) 探测器及用于智能划分能区的探测系统和方法
Mohammadian-Behbahani Non-iterative pulse tail extrapolation algorithms for correcting nuclear pulse pile-up
BR112016013655B1 (pt) Método e aparelho para resolução de sinais de dados

Legal Events

Date Code Title Description
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20170420

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

AX Request for extension of the european patent

Extension state: BA ME

DAV Request for validation of the european patent (deleted)
DAX Request for extension of the european patent (deleted)
REG Reference to a national code

Ref country code: DE

Ref legal event code: R079

Ref document number: 602015068798

Country of ref document: DE

Free format text: PREVIOUS MAIN CLASS: G01N0023080000

Ipc: G01N0023040000

A4 Supplementary search report drawn up and despatched

Effective date: 20180625

RIC1 Information provided on ipc code assigned before grant

Ipc: G01N 23/04 20060101AFI20180619BHEP

Ipc: G01N 23/22 20060101ALI20180619BHEP

Ipc: G01V 5/00 20060101ALI20180619BHEP

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: EXAMINATION IS IN PROGRESS

17Q First examination report despatched

Effective date: 20200305

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: GRANT OF PATENT IS INTENDED

INTG Intention to grant announced

Effective date: 20210115

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE PATENT HAS BEEN GRANTED

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

REG Reference to a national code

Ref country code: GB

Ref legal event code: FG4D

REG Reference to a national code

Ref country code: CH

Ref legal event code: EP

REG Reference to a national code

Ref country code: DE

Ref legal event code: R096

Ref document number: 602015068798

Country of ref document: DE

REG Reference to a national code

Ref country code: AT

Ref legal event code: REF

Ref document number: 1387618

Country of ref document: AT

Kind code of ref document: T

Effective date: 20210515

REG Reference to a national code

Ref country code: IE

Ref legal event code: FG4D

REG Reference to a national code

Ref country code: FI

Ref legal event code: FGE

REG Reference to a national code

Ref country code: NL

Ref legal event code: FP

REG Reference to a national code

Ref country code: DK

Ref legal event code: T3

Effective date: 20210714

REG Reference to a national code

Ref country code: NO

Ref legal event code: T2

Effective date: 20210428

REG Reference to a national code

Ref country code: SE

Ref legal event code: TRGR

REG Reference to a national code

Ref country code: LT

Ref legal event code: MG9D

REG Reference to a national code

Ref country code: AT

Ref legal event code: MK05

Ref document number: 1387618

Country of ref document: AT

Kind code of ref document: T

Effective date: 20210428

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: LT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: AT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: BG

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210728

Ref country code: HR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

REG Reference to a national code

Ref country code: ES

Ref legal event code: FG2A

Ref document number: 2877564

Country of ref document: ES

Kind code of ref document: T3

Effective date: 20211117

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: PL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: PT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210830

Ref country code: RS

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: GR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210729

Ref country code: LV

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: IS

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210828

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SK

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: SM

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: EE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: CZ

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

Ref country code: RO

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

REG Reference to a national code

Ref country code: DE

Ref legal event code: R097

Ref document number: 602015068798

Country of ref document: DE

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed

Effective date: 20220131

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: IS

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210828

Ref country code: AL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: MC

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: LU

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20211128

Ref country code: BE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20211130

REG Reference to a national code

Ref country code: BE

Ref legal event code: MM

Effective date: 20211130

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: ES

Payment date: 20230125

Year of fee payment: 8

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: HU

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT; INVALID AB INITIO

Effective date: 20151128

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: CY

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20210428

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: NL

Payment date: 20231120

Year of fee payment: 9

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20231123

Year of fee payment: 9

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: SE

Payment date: 20231120

Year of fee payment: 9

Ref country code: NO

Payment date: 20231124

Year of fee payment: 9

Ref country code: IT

Payment date: 20231124

Year of fee payment: 9

Ref country code: IE

Payment date: 20231121

Year of fee payment: 9

Ref country code: FR

Payment date: 20231120

Year of fee payment: 9

Ref country code: FI

Payment date: 20231121

Year of fee payment: 9

Ref country code: DK

Payment date: 20231124

Year of fee payment: 9

Ref country code: DE

Payment date: 20231121

Year of fee payment: 9

Ref country code: CH

Payment date: 20231201

Year of fee payment: 9

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: ES

Payment date: 20240129

Year of fee payment: 9